Identifying the Common Functions of Genes Linked to Autism Spectrum Disorder (ASD)

Overview

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder with a significant genetic contribution. Studies have identified hundreds of genes with evidence of involvement in ASD when mutated (“ASD genes”). Despite the vast number of ASD risk genes identified, our knowledge of these genes isn’t sufficient to use genetic information in ASD treatment or diagnosis in most cases. Due to the large number and diversity of ASD genes, I hypothesized that ASD genes share common functions in mammalian development. The functions of a gene can be determined by deleting (or “knocking out”) the gene in mice and observing the phenotype(s) that result. This study aimed to advance our understanding of ASD genetics by revealing common functions of ASD genes. T o test my hypothesis, I performed bioinformatic and statistical analysis on a large collection of phenotyping data from the International Mouse Phenotyping Consortium (IMPC), which has cataloged 55,337 phenotypes that resulted from 7,093 gene knockouts in mice, including 395 orthologs of human ASD genes. I found that several phenotypes are significantly enriched among ASD gene knockouts relative to non-ASD gene knockouts. These phenotypes include embryonic lethality, increased thigmotaxis, hyperactivity, increased anxiety-related responses, decreased exploration in new environments, abnormal contextual conditioning behavior, increased circulating iron level, and increased lean body mass. The phenotypes identified here point to underlying developmental pathways and processes likely involved in the etiology of ASD and should be further studied. Furthermore, these results identify phenotypes that could be used to evaluate the validity of ASD mouse models used. Novel Brain-Computer Interface for Binary Communication in Non-Verbal Patients Through Motor Imagery Julian Varga Gwinnett School of Mathematics, Science and Technology, Lawrenceville, GA Speech and language are essential for social interactions and communication of medical problems. Patients with diseases such as ALS, locked-in syndrome, aphasia, and patients recovering from strokes are typically unable to communicate their basic needs to their doctors, resulting in low quality of life. The current best solution to this is Speech Generating Devices, which can often cost upwards of $15,000 and have an incredibly difficult learning curve. This study aims to design an easy-to-use and cost-effective brain-computer interface (BCI) using only 4 electrodes to aid those with nonverbal communication disorders. The BCI detects electrical signals related to left and right arm motor functions, corresponding to ‘yes’ or ‘no’ responses, allowing patients to imagine moving a certain arm to communicate this binary response without moving or speaking. The researchers developed a new novel preprocessing algorithm, the JMS Breakdown Algorithm, and found that the algorithm significantly outperforms other existing preprocessing algorithms in reducing classification errors and increasing information transfer rates. The novel use of an attention layer in the machine learning model also boosted classification accuracy by 4%, giving the machine a final classification error of 4.62% when tested at a set frequency of 11 Hz using both the JMS Breakdown Algorithm and the CNN-RNN with an attention layer, statistically lower than previous 4-electrode BCIs. The methods used in this paper can also be applied to BCIs of any use for improved accuracy, specifically the JMS Breakdown Algorithm, further enhancing the field of BCIs overall. GREATER WASHINGTON, D.C.

Competition history

• JSHS 2023 Category not listed

Resources

• National JSHS abstract book ↗

Related projects

Closest projects by meaning, across every fair and year in the corpus.